Partitioned Path Loss Models Based on Coefficient of Determination

A path loss model is a fundamental method to roughly predict radio propagation characteristics. However, most conventional models do not consider the anisotropy of radio propagation. If geographical conditions, such as the number of buildings, are irregularly change, the radio propagation characteristics may significantly fluctuate. This paper proposes the partitioned path loss models to precisely estimate the radio propagation. The proposed method first collects the instantaneous received signal power via a measurement campaign. Then, measured datasets are transformed to the polar coordinate. Thus, several azimuth regions are created centered on the transmitter, and a path loss model is estimated in each azimuth region. Through the performance evaluation, we can confirm that the proposed method accurately estimates the radio propagation compared to the single path loss model.